two rare oxytropis species (fabaceae) endemic to the pirin

335

8• Phytol. Balcan. 13(3) • 2007

PHYTOLOGIA BALCANICA 13 (3): 335 –346, Sofia, 2007

Two rare Oxytropis species (Fabaceae) endemic to the Pirin Mts, Bulgaria

Ekaterina Kozuharova1, A. John Richards2, Marie Hale3& Kirsten Wolff2

1 Department of Pharmacognosy, Faculty of Pharmacy, 2 Dunav St., 1000 Sofia, Bulgaria, email: [email protected] (corresponding author)

2 School of Biology , Ridley Building, University of Newcastle NE1 7RU, UK3 School of Biol. Sciences University of Canterbury Private Bag, 4800 Christchurch, 8140,

New Zealand

Received: March 27, 2007 ▷ Accepted: October 05, 2007

Abstract. Using morphological characters and molecular techniques, including RAPDs, ITS and trnL sequences, we compared O. urumovii and O. kozhuharovii, two alpine Oxytropis endemics to the North Bulgarian Pirin marble, with neighbouring populations of the widespread O. campestris, O. halleri and O. dinarica. Oxytropis urumovii is a very distinct diploid species which might be ancestral to this group and could be regarded as a palaeoendemic. The tetraploid O. kozhuharovii is most closely related to O. prenja from the Dinaric Alps, Bosnia-Herzegovina, but is a larger plant with a different facies and indumentum. It is possible that it has evolved as an allotetraploid derivative of O. urumovii and O. halleri. It is also possible that the circumpolar hexaploid O. campestris has evolved as an allohexaploid derivative of the diploid O. urumovii and a tetraploid from the Balkans, such as O. kozhuharovii. We present a key to summarise the differences between the Balkan Oxytropis species. Although about thousand mature individuals of O. kozhuharovii survive in the wild, the population occupies less than 200 m2, so this species is considered to be Critically Endangered.

Key words: ITS sequence, Oxytropis kozhuharovii, O. urumovii, phylogeny, RAPD

Introduction

Although isolated by low ground from the Rila Mts to the north and the Central Rhodope Mts to the south, the Pirin Mts of Southwest Bulgaria neverthe-less lie within a region of considerable topographical and geological complexity, known as the Rhodopean Supergroup. Even within the Pirin Mts, five distinct anticlines running roughly south-west to north-east can be identified, culminating in the highest peak, Vihren, which at 2914 m is the third highest moun-tain in the Balkans. The Pirin Mts result from recent Neogenic-Quaternary uplifts, although the rocks in-volved are extremely complex, dating from a wide variety of geological periods, and have mostly been metamorphically modified. Much of the northern Pirin Mts (Razlozhki and Sinanitsa anticlines) is mar-

ble (Zagorchev 1995). Velchev & Kenderova (1994) differentiated three glaciations of the mountain on the basis of thermoluminescence analyses of correlative deposits from the western foot of Pirin: Mindell, Riss and Wurm.

Perhaps as a result of this recent origin and glacial history, as well as the complex geology and topogra-phy, the Pirin Mts host more than 20 species endemic to the Pirin marble, according to Velčev & al. (1992). This diversity of local endemics is comparable with that of the better known Greek mountain Olimbos which lies 200 km to the south and is of very simi-lar bulk and altitude. Amongst these Pirin endemics are two alpine species of Oxytropis: O. urumovii Jav. and O. kozhuharovii Pavlova, Dimitrov & Nikolova. A third alpine species, O. campestris (L.) DC. subsp. campestris, also occurs in these mountains (Fig. 1).

336 Kozuharova, E. & al. • Two rare endemic Oxytropis species

Fig. 1. Distribution of the acaulescent Oxytropis species in the Balkans (compilation aft er Hayek 1927; Leins & Merxmüller 1968; Diklich 1972; Kožuharov 1976; Chrtek & Chrtkova 1983; Strid 1986).

Yellow fl owers O. dinarica (Murb.) Wettst. subsp. velebitica Chrtek & Chrtkova: Mt Velebit, 1400–1600 m; O. dinarica subsp. dinarica: Mt Velez ,1600–1800–2300 m; O. dinarica subsp. weberi Chrtek & Chrtkova: Mt Korab, 2000 m, Popova Šapka, Šar Planina, NW Macedonia (2n = 16); O. urumovii Jáv. (2n = 16), Pirin marbles, above 2500 m; O. campestris (L.) DC. subsp. campestris (2n = 48), Pirin and Rila marbles, at about 2500 m, Prizren (Kosovo), Carpathians.

Purple fl owers O. kozhuharovii Pavlova, Dimitrov & Nikolova (2n = 32), Pirin marbles, above 2500 m; O. prenja (Beck) Beck, 2350 m shist, Gramos and mountains of the W part of the Balkan Peninsula; O. halleri subsp. korabensis (Kümmerle & Jav.) Chrtek & Chrtková, Mt Korab, Prizren (Kosovo); O. halleri Bunge ex Koch, Carpathians and Slovenia (2n = 32); O. purpurea (Bald.) Markgr. (2n = 16), limestone ridges 1900–2800, Olimbos, Albania, Macedonia

The fourth Bulgarian Oxytropis, O. pilosa (L.) DC. is a subalpine species which is widespread in East Europe but occurs only in a few low-altitude mountains in West and East Bulgaria.

The genus Oxytropis has about 300 species distrib-uted through temperate, montane and boreal regions of the northern hemisphere. It is closely related to an even larger and more widespread genus, Astragalus (2500 spe-cies), from which it is distinguished only by a mucronate tooth on the abaxial side of the keel. It seems possible that Oxytropis is an artificial grouping; for instance, some spe-cies in both genera are ‘locoweeds’, selenium accumula-tors which are harmful to stock. However, Wojciechowski & al. (1993) and Sanderson & Wojciechowski (1996)

show that three species of Oxytropis, including O. camp-estris, are sisters to a much larger sample of Astragalus which forms a single clade, but studies involving a larger number of Oxytropis species are awaited.

An interesting hypothesis on circumpolar specia-tion in the genus Oxytropis is offered by Yurtsev (1999). Only three genera of legumes (Oxytropis, Astragalus, Hedysarum) are typical of the Arctic, and only Oxytropis has undergone intense speciation there. Only a few sub-divisions of the genus Oxytropis have contributed to this Arctic florogenesis and speciation. In the O. campes-tris “superaggregate” the lowest ploidy level in Eurasian species and races is 2n = 48, whereas in western North America and Far-East Asia there are taxa with both 2n =

337Phytol. Balcan. 13(3) • Sofia • 2007

Table 1. Sites from which Oxytropis were collected in the wild. Datum WGS 1984, UTM projection.No Species Site Altitude

mLongitude

E (W)Latitude

Ncollection

date3 O. campestris Vihren chalet to Kabata 2410 23°24´10.3˝ 41°45´38.1˝ 08.07.20015 O. campestris Kabata 2640 23°23´50.6˝ 41°45´38.5˝ 08.07.20017 O. campestris Kazan I, Vihren 2605 23°24´11.3˝ 41°46´17.2˝ 08.7.20018 O. urumovii Kazan I, Vihren 2605 23°24´10.8˝ 41°46´17.3˝ 08.07.200116 O. kozhuharovii Yellow rocks, Okadenski cirque 2170 23°20´45.5˝ 41°49´12.3˝ 28.07.200119 O. campestris Razlozhki Souhodol 2570 23°21´41.6˝ 41°47´52.7˝ 29.07.200120 O. urumovii Razlozhki Souhodol 2570 23°21´41.6˝ 41°47´52.7˝ 29.07.200121 O. urumovii Kazan II, Vihren 2251 23°24´50.9˝ 41°46´18.2˝ 31.07.200122 O. campestris Vihren chalet to Kabata 2200 23°24´54.3˝ 41°45´26.8˝ 01.08.2001

O. halleri Invernaver, UK 5 04°13´40˝W 58°31´48˝ 08.08.2002O. halleri Strathy, UK 25 03°59´55˝W 58°34´20˝ 08.08.2002O. dinarica subsp. weberi

Meadows, Popova Šapka, Šar Planina, NW Macedonia

06.1996 (A. Novotna)

16 and 32, which are not always morphologically dis-tinct: e.g. O. gorodkovii in the easternmost Chukotka Peninsula, (Yurtsev 1986, 1988, 1994, 1997, 1999; Yurtsev & al. 1978, 1994). At the same time, morphometric anal-yses reveal significant differences between some North American taxa – the tetraploid O. campestris var. chart-acea, and the hexaploid O. campestris var. johannensis (Chung & al. 2004). In other cases, polyploidy has ac-companied and consolidated evolutionary adaptation to contrasting cold treeless environments (as in O. sver-drupii 2n = 48, O. wrangelii 2n = 64; or O. viscida 2n = 16 and 32, O. middendorffii 2n = 48) (Yurtsev 1986, 1988, 1994, 1997, 1999; Yurtsev & al. 1978, 1994). In contrast, neither molecular nor morphometric data support the differentiation of the Eastern Alpine endemic O. camp-estris subsp. tiroliensis from the widespread O. c. subsp. campestris (Schönswetter & al. 2004). Onobrychis camp-estris s.str. is a glacial relict on the Balkan peninsula con-nected to the central European mountains and especially to the Alps (Andreev 1989). The evolution of the whole series Sordidae is based on auto- and aloploidy (Küpfer 1974; Andreev 1989). As a whole the taxonomy of O. campestris s. latiss. remains unsolved.

There are several species of Oxytropis in the Balkans (Fig. 1, Hayek 1927; Leins & Merxmüller 1968; Diklich 1972; Kožuharov 1976; Chrtek & Chrtkova 1983; Strid 1986). In this study we focus on the acaulescent spe-cies with semi-bilocular and almost completely biloc-ular legumes.

The aims of the present study were to use morpho-logical and molecular evidence to investigate the dis-tinctness and add “a brick in the wall” in the efforts to

elucidate the possible evolutionary origins of O. uru-movii and O. kozhuharovii. Both taxa are strongly lo-calised in the wild, and if their specific rank can be justified, their conservation becomes a matter of some urgency. Unfortunately, it has not been possible to in-clude O. prenja (Beck) Beck in the study.

Materials and methods

In late July 2001, the authors visited 10 sites with Oxytropis in the northern Pirin Mts: six with O. camp-estris, four with O. urumovii and the only known site for O. kozhuharovii. Those sites used in the analysis are listed in Table 1. At each site, samples of leaves, flowers and fruits were taken from a minimum of five individuals for morphological examination, and leaf material was dried with silica gel for later DNA ex-traction. In August 2002, material of O. halleri sub-sp. halleri was collected from five individuals at each of the two sites on the northern coast of Scotland, and in July 2003, material of O. dinarica subsp. weberi was collected from Northwest Macedonia (Table 1).

A minimum number of two flowers per plant and a minimum of five plants per site were rehydrated in warm dilute alcohol, dissected into components, dehy-drated and mounted on a sheet. The sheets were digi-talized (images scanned at 1:1). A total of 14 characters were measured on flowers and a further three measure-ments were made of leaves (Table 2, Fig. 5). The flow-er characters were measured digitally using Adobe Photoshop 5.0. Measurements of the leaves were made


Table 2. Characters recorded in morphological analysis.

Length of bract, cmg Maximum length of calyx, cmi Maximum length of calyx tube, cmh Maximum width of calyx, cmg-i Maximum length of calyx teeth, cmb Maximum length of standard petal, cmb´ Length of beak of the standard petal, cma Maximum width of standard petal cma´ Width of the standard petal close to the top, cma˝ Width of the standard petal close to the beak, cmd Length of keel petals, cmc Maximum width of keel petals, cme Length of beak to keel, cm (the narrower base to the keel)e´ Length of keel tooth, cmf Maximum length of wing petal, cmf´ Maximum width of wing petal, cmf˝ Length of beak to wing, cm

Length of rhachis to leaf, cmTotal number of leaflets/leaf (leaflets per side = [n/2] – 1)Length/width median leaflet on a leaf

Fig. 5. Characters recorded in morphological analysis. Th e legend is in Table 2.

using a pocket micrometer. Descriptive statistics were calculated for individuals within populations, and us-ing this data, for populations for each species. To an-alyse morphological differences between taxa, average readings per population were used for O. campestris, but measurements were only taken for single popula-tions of O. urumovii and O. kozhuharovii, and for these species mean readings per individual were computed.

Inspection showed that six floral characters were diffferentiated between at least two of the species ex-amined, and these were used in a principal compo-nents analysis (PCA), using a covariance matrix. Two of the three vegetative characters examined differenti-ated at least two of the taxa.

DNA was extracted using a CTAB with chloro-form method (Weising & al. 1995) from five indi-

viduals per sampling location. We amplified a to-tal of 46 polymorphic RAPD fragments using four primers (Operon Technologies): OPA8 (12 polymor-phic bands), OPA12 (15 polymorphic bands), OPA6 (11 polymorphic bands), and OPA4 (8 polymorphic bands) in 25μl reactions. Each reaction consisted of 1 x Taq buffer (16 mM (NH4)2SO4, 67 mM Tris-HCl, 0.01 % Tween-20), 2.0 mM MgCl2, 0.08 mM each dNTP, 0.2 μM primer, 1.0U Taq (Bioline) and 0.5 μl template DNA. The reaction cycle was 94 °C for 3 min, then 45 cycles of 94 °C for 30 sec, anneal-ing temperature for 30 sec, 72 °C for 1.5 min, with a final extension of 72 °C for 4 min. Annealing tem-peratures were 40 °C for OPA8 and OPA12, 36 °C for OPA6 and OPA4. All reactions were performed in a PTC-100TM thermocycler (MJ research). PCR prod-ucts were visualised on 1.4 % agarose gels stained with ethidium bromide. We regarded the presence/absence state of RAPD bands, comparable between species and individuals, as a data matrix of bi-state characters varying between each individual. . We viewed the relationships between these as a sin-gle linkage dendrogram, based on a simple euclid-ian distance matrix, using cluster analysis available within the Principle Components Analysis program in the MINITAB 14 statistical package.

Nei’s genetic distance between sampling locations was calculated from the RAPD presence/absence da-ta using RAPDDIST 1.0 (Black 1995), with 1000 boot-strap replications. A consensus neighbour-joining tree was calculated from the bootstrap replicate Nei’s distance matrices using PHYLIP 3.57c (Felsenstein 1993), and was visualised using TREEVIEW 1.6.1 (Page 1996).

The chloroplast TrnL intron was amplified for 24 in-dividuals (9 O. campestris, 8 O. urumovii, and 7 O. ko -zhu harovii) using primers TrnL-c and TrnL-d (Taberlet & al. 1991) in 25 μl reactions containing 1x Taq buffer, 2.0 mM MgCl2, 0.2 mM each dNTP, 0.2 μM each prim-er, 1.0U Taq (Bioline), and 0.5 μl template DNA. The reaction cycle was 35 cycles of 93 °C for 1 min, 50 °C for 1 min, 72 °C for 2 min. The internal transcribed spac-er region (ITS) was also amplified for each of these 24 individuals using the primers ITS4 and ITS5 (White & al. 1990) and the same PCR component concentrations as for the TrnL intron amplification. The reaction cy-cle for ITS was 94 °C for 5 min followed by 40 cycles of 94 °C for 30 sec, 50 °C for 30 sec, 72 °C for 2 min, with a final extension of 72 °C for 4 min.


0.8

0.9

1

1,1

1.2

1.3

1.4

0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

calyx teeth length cm

keel

leng

th cm

urumovii

dinarica

campestris

kozhuharoviiFig. 6. Mean values in cm of calyx tooth length and keel length for four species of Oxytropis. Readings for O. urumovii and O. kozhuharovii are means of individuals from single populations. Readings for O. campes-tris and O. dinarica are means of fi ve individuals for a population.

PCR products were sequenced directly. All PCR products were purified using QIAquick® PCR Purification Kits (Qiagen). Purified PCR prod-ucts were then sequenced using BigDye Terminator Cycle Sequencing chemistry (Applied Biosystems) following manufacturer’s recommended conditions, and sequences detected on an ABI 310 Prism® auto-mated sequencer (Applied Biosystems). Sequences were aligned and edited using ProSequence (Filatov 2002).

Most species in this study had been character-ised cytologically previously, but the chromosome number of O. dinarica was unknown. Material was collected by A. Novotna from the meadows of Popova Šapka, Šar Planina, NW Macedonia in July 1996 (voucher kept in personal collection in the Faculty of Pharmacy, Sofia). The haploid chromo-some number in the pollen grains was studied by the acetocarmine method suggested by Heywood (1967) and Taschetto & Pagliarini (2004) on seven buds at different stage and two flowers. As a control we used 10 flowers and buds of O. campestris from the Pirin and Rila Mts, with a known chromosome number (n = 24).

Results

Morphological analysis

Using pairs of flower characters, the only combina-tion which completely separated the species was calyx teeth length and keel length (Fig. 6), although O. uru-movii also has longer bracts and a longer calyx than

the other species (Fig. 7), and is undoubtedly the most distinct of the four species studied. Generally, the yel-low-flowered O. dinarica falls closest to the purple-flowered O. kozhuharovii morphologically (Fig. 8). Oxytropis kozhuharovii has narrower and less numer-ous leaflets than most O. campestris (Fig. 4). In gener-al, the hexaploid O. campestris tends to fall between the diploid O. urumovii and the tetraploid O. kozhu-harovii, morphologically.

The haploid chromosome number in the pollen grains of Oxytropis dinarica was counted as n = 8.

Molecular analysis

Neither ITS nor TrnL were particularly variable. For both markers, only two polymorphic sites were de-tected out of 452 base pairs sequenced (ITS) and 454 base pairs (TrnL). A neighbour-joining tree of the TrnL polymorphisms (two insertion/deletions) dif-ferentiated O. urumovii from the other species, but did not differentiate O. kozhuharovii from O. camp-estris. The ITS polymorphisms (two base substitu-tions) did not differentiate any of the species. Only the RAPD data provided enough resolution to ade-quately examine variation between the sampling lo-cations.

The results from the neighbour-joining un-rooted tree based on RAPD presence/absence data (Fig. 9) and from the principal component dendro-gram analysis of the RAPD band matrix (Fig. 10) are very similar. In both analyses, the four species examined remained distinct, with good bootstrap support in the neighbour joining analysis of Nei’s genetic distance.


0.6

0.7

0.8

0.9

1

1.1

1.2

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

bract length cm

caly

x le

ngth

cm

urumovii

dinarica

campestris

kozhuharovii

Fig. 7. Mean values in cm of bract length and calyx length for four species of Oxytropis. Readings for O. urumovii and O. kozhuharovii are means of individuals from single populations. Readings for O. campes-tris and O. dinarica are means of fi ve individuals for a population.

Fig. 8. Principal components analysis fi rst and second vectors aft er multivariate analysis for six fl oral characters in four species of Oxytro-pis. Readings for O. urumovii and O. kozhuharovii are means of individuals from single populations. Readings for O. campestris and O. dinarica are means of fi ve individuals for a population.

Fig. 4. Mean values for fi ve leaves of leafl et shape (length/width) and total leafl et number per leaf for six individuals of O. kozhuharovii and 27 individuals of O. campestris.

Oxytropis flower morphology PCA

-500

-400

-300

-200

-100

0

100

200

300

400

500

-310 -290 -270 -250 -230 -210 -190

dinarica

campestris

urumovii

kozhuharovii

Oxytropis, leaflet number versus average leaflet shape

1

1.5

2

2.5

3

3.5

4

4.5

10 12 14 16 18 20 22 24 26 28 30

leaflet number

leaf

let s

hape

kozhuharovii


Fig. 9. Nei’s genetic distance between sampling locations calculated from the RAPD presence/absence data with 1000 bootstrap replications. A consensus neighbour-joining tree was calcu-lated from the bootstrap replicate Nei’s distance matrices and visualised using TREEVIEW.

Fig. 10. Dendrogram of clus-ter analysis in ten populations of Oxytropis classified in four species, based on presence or absence of 46 RAPD bands.

Discussion

Oxytropis urumovii is a distinctive species compared on first sight to the other Balkan Oxytropis acaules-cent species (with a semi-bilocular or almost com-pletely bilocular legume, Appendix 1, Figs. 1 & 2), The whole plant, including the fruit, is rather dense-ly covered with long (2–3 mm) patent whitish hairs, which contrast with the scape and calyx, the green col-our of which is often suffused with black. Amongst the European species of Oxytropis, O. urumovii is di-agnosed by being acaulescent, having leaves with about eight pairs of leaflets, nearly free stipules (ad-nate to the petiole for no more than one quarter), ca-lyx-teeth which are much shorter than the tube, yel-lowish flowers and an ovoid semi-bilocular legume

(Fig. 2, Leins & Merxmüller 1968; Kožuharov 1976). It is diploid with 2n = 16 (Kruscheva 1986; Pavlova 1996). The karyotype is symmetrical, consisting of 2n = 4m + 12sm = 16 small and medium size chro-mosomes. (Pavlova 1996). In addition, the author notes that endopolyploidy (2n = 48) is also observed and that is how she interprets the previous report of a hexaploid number (Andreev 1981). It is immediate-ly distinguished from O. campestris by the much long-er, denser and generally patent indumentum (usually sparse and semi-appressed in O. campestris), flowers that turn reddish rather than blackish at the tip, nearly free stipules (in O. campestris stipules are adnate to the petiole for one-third to three-quarters of their length), and a smaller legume (see also Leins & Merxmüller 1968; Kožuharov 1976). In general, it has a dwarfer

O. urumovii2n = 16

O. kozhuharovii2n = 32

O. halleri2n = 32

O. campestris2n = 48

(O. hallerii) 0.1

22 (O. campestris)

16(O. kozhuharovii)

StrathyBettyhill(O. hallerii)

21(O. urumovii)

8(O. urumovii)

20 (O. urumovii)19 (O. campestris)

7 (O. campestris)5 (O. campestris)

100

52

63

97

80 32 48


Fig. 2. Legumes of O. uru-movii, O. kozhuharovii and O. campestris – views of the central vein, accretion, and (below) transverse sections.

1. Plant caulescent, stem >20 cm; calyx teeth >calyx tube; flowers yellow (widespread subalpine) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . O. pilosa

1*. Acaulescent alpine; calyx teeth <half length of tube; flowers yellow or purple. . . . . . . . . . . . . . . . 2

2. Calyx teeth <25 % length of calyx tube; hairs on fruit 2–3 mm, dense; flowers yellow (Croatia to Albania and Macedonia, three subspecies). . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .O. dinarica

2*. Calyx teeth >25 % length of calyx tube; hairs on fruit usually less than 1 mm (except occasionally O. urumovii); flowers yellow or purple . . . . . . . . 3

3. Fruit bilocular, with a septum in the dorsal valve; stipules with (1-) 2 or more veins; leaf indumen-tum velutinous (the hairs would be less than 0.5 mm long, although it is their soft dense quality that is important); flowers purple (widespread, but only in Mt Korab in the Balkans, e.g. widespread subsp. halleri 2n = 32, subsp. velutina – 2n = 16 re-stricted to the Alps, but in Mt Korab subsp. kora-bensis . . . . . . . . . . . . . . . . . . . . . . . . . . . . O. halleri

3*. Fruit semi-bilocular, lacking a septum in the dor-sal valve; stipules one-veined; leaf indumentum usually pilose or sericeous but not velutinous; flowers yellow or purple . . . . . . . . . . . . . . . . . . . . 4

4. Stipules free; calyx <1 cm; indumentum dense; in-florescence peduncle somewhat exceed the sub-tending leaf; raceme 4–12-flowered, flowers cher-

ry-red; peduncle and legume with long, patent hairs (N Greece, Albania, SW Macedonia). . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . O. purpurea

4*. Stipules adnate to petiole. . . . . . . . . . . . . . . . . . . . 55. Stipules adnate to petiole for 25 % or less; calyx

>1 cm and keel >1.2 cm in length; flowers yellow and indumentum dense; inflorescence peduncle about as long as subtending leaf – shorter or equal, seldom longer; throughout (N Pirin). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . O. urumovii

5. Stipules adnate to petiole for half length or more; calyx <1 cm, keel <1.2 cm; flowers yellow or pur-ple; indumentum sparse except occasionally on calyx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

6. Hairs on calyx long and dense, especially at apex, exceeding calyx teeth in length; inflorescences ex-ceeding leaves; flowers purple (one site in N Pirin . . . . . . . . . . . . . . . . . . . . . . . . . . . . O. kozhuharovii

6*. Calyx with short sparse indumentum, or nearly glabrous; inflorescences not exceeding leaves (pe-duncle about as long as subtending leaf) flowers yellow or purple . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

7. Leaflets usually >9 pairs; flowers yellow (wide-spread alpine . . . . . . . . . . . . . . . . . . . O. campestris

7*. Leaflets 6–8 pairs; flowers purple; raceme 1–4-flowered, peduncle and legume with short, ap-pressed hairs (Dinaric Alps, Bosnia-Herzegovina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .O. prenja

Appendix 1***

Key to Balkan species of Oxytropis with semi-bilocular or almost completely bilocular legume.


Fig. 3. O. kozhuharovii, illustrating the narrow leafl ets and the diagnostic long hairs at the apex of the calyx.

and stiffer habit than the rather lax and sprawling O. campestris. According to Pavlova (1996), the karyo-type of the population O. campestris in the Pirin Mts consists of 2n = 24m + 20sm + 4st = 48, and the chro-mosomes are of medium size.

Leins & Merxmüller (1968) included O. camp-estris subsp. dinarica Murb. within O. urumovii, but Chrtek & Chrtkova (1983) reinstated the combination of Wettstein (1892) – O. dinarica (Murb.) Wettst., and described two new subspecies for this taxon, which occurs from Croatia (subsp. velebitica) to Bosnia and Montenegro (subsp. dinarica) and to Macedonia and Albania (subsp. weberi) (Fig. 1). In general, O. dinarica lies between O. campestris and O. urumovii, not least in general habit. It resembles the former species by the stipules which are adnate to the petiole for half their length, but the indumentum resembles more O. uru-movii, except on the fruit on which the short (<1 mm) hairs are similar to those of O. campestris. The calyx of O. dinarica is diagnostic, for the calyx teeth are on-ly about one-fifth the length of the calyx tube, but in the other two species they are more than one-quarter the length of the tube. In all three of these species the flowers are yellowish. O. dinarica is reported as dip-loid (2n = 16) for the first time in this paper. It is clear-ly a distinct species, so that O. urumovii s. s. can be re-garded as being endemic to the Bulgarian Pirin Mts.

Unlike the species mentioned previously, the oth-er Pirin endemic Oxytropis, O. kozhuharovii is a tetra-ploid with 2n = 32 (Pavlova & al. 1999). It is only known from a single locality from the northernmost part of the Pirin, the Yavorov anticline, where it was discovered and described very recently (Pavlova & al. 1999). Oxytropis kozhuharovii is the only Bulgarian Oxytropis with blue or purple flowers. However, at least one and possibly two purple-flowered spe-cies occur in the region around the border between Macedonia and Albania, some 220 km to the west, no-tably Mt Korab (Fig. 1). Here an endemic subspecies of the widespread species O. halleri Bunge ex Koch, O. h. subsp. korabensis (Kummerle & Jav.) Chrtek & Chrtkova, occurs (Fig. 1). As discussed by Pavlova & al. (1999), O. kozhuharovii closely resembles mem-bers of the O. halleri complex, but differs in the struc-ture of the legume which, like that of O. campestris and O. urumovii, lacks a septum in the dorsal valve (semi-bilocular). All subspecies of O. halleri have a septum in the dorsal valve, although that in subsp. ko-rabensis is narrow and poorly developed. Stipules are

another feature in which O. kozhuharovii resembles more the O. campestris complex than O. halleri. Most forms of O. halleri have stipules with several longitu-dinal veins, whereas those of O. kozhuharovii are sin-gle-veined. However, the reduced stipules of O. halleri subsp. korabensis have only 1–2 veins. We would add that the indumentum of the O. halleri complex tends to be velutinous, but that in O. kozhuharovii is long-er and patent.

In our opinion, the most distinctive feature of O. kozhuharovii is the indumentum of the calyx, partic-ularly the apex, which is very densely covered with long white hairs, which equal or exceed the calyx teeth (Fig. 3). This contrasts strikingly with the rest of the plant in which the patent indumentum is long but sparse and inconspicuous.

Although it resembles O. halleri superficially, it seems likely that O. kozhuharovii is in fact more closely related to O. prenja which also has semi-bilocular legumes. From an examination of the her-barium material at K and E, it is clear that the lat-ter is a dwarfer plant in which the scape exceeds the rather prostrate foliage. Decisively, it is scarcely hairy

9• Phytol. Balcan. 13(3) • 2007


even on the calyx. It is also said to have fewer pairs of leaflets per leaf (usually 6–7 rather than 10–11 for O. kozhuharovii), but this disagrees with our find-ings for the latter species (Fig. 4). Also, according to Leins & Merxmüller (1968), O. prenja is said to have a longer (15–19 mm) standard which is emarginate at the apex, and a shorter keel tooth (‘beak’) of on-ly 0.5 mm. However, although Fig. 4 in Pavlova & al. (1999) appears to show an emarginate standard in O. prenja, the apex of the standard in this species is said in Table 1 to be ‘convex’, and the standard illus-trated does not appear to differ in length from that of O. kozhuharovii at about 1.3 mm. Also, although Pavlova & al. (1999) state that the length of the keel tooth in the latter species is ‘circa 1 mm’ (diagnosis) or ‘1–1.5 mm’ (final line), the present study found that it varied between 0.3 and 0.8 mm, and did not differ from that of O. prenja in this regard.

Oxytropis prenja was ignored by Chrtek & Chrtkova (1983) and seems to be little known out-side its native area. The description in Leins & Merxmüller (1968) would not distinguish it from O. kozhuharovii. However, according to a recent visi-tor, its general appearence is unlike that of the lat-ter species (fid. Anonymous reviewer). Its image has twice appeared on postage stamps for Bosnia-Herzegovina, as a painting (1997) and a photograph (2003). Accompanying website texts claim it as an endemic of the Dinaric Mts, Prenj Planina, Čvrsnica, and Vran, where it occurs in limestone crevices at al-titudes of 1900–2228 m. This taxon has also been said to occur in Albania and Macedonia (Greuter 1989), where it may have been confused with O. halleri sub-sp. korabensis.

We conclude that each of the localised Balkan en-demic Oxytropis are distinct species. In our view, they are best differentiated as shown in the key (Appendix 1).

The molecular analysis revealed the following: O. urumovii proved to be the most distinct species of the four. As it is diploid, it could be argued that O. uru-movii is ancestral to this group and a palaeoendemic element of the flora of the Pirin Mts.

Compared to Bulgarian O. campestris and Scottish O. halleri, O. kozhuharovii is about equally related to each. Both analyses would be consistent with the hy-pothesis that O. campestris could have arisen from the allohexaploid union between O. urumovii and O. kozhuharovii, thus agreeing with the morphological

data (unfortunately, material of O. dinarica was re-ceived too late to be included in this analysis).

In addition, the molecular evidence suggests that O. kozhuharovii could be the allotetraploid deriva-tive of the sympatric diploid endemic O. urumov-ii and the widespread diploid (and tetraploid) O. halleri, the nearest populations of which occur today in Mt Korab, some 220 km away. Today, this disjunc-tion seems significant, but during the Late Glacial stadials mountain steppe vegetation (Artemisia – Chenopodiaceae – Poaceae) in which they occur dom-inated at high and mid-altitudes (Wijmstra 1969; Florschütz & al. 1971; Bottema 1974, 1979; Lang 1994; Tzedakis 1994, 1999; Bozilova 1996; Bozilova & al. 1989; Bozilova & Tonkov 2000; Tonkov & al. 2002, 2006; etc). At that time species that now occur at a higher altitude were more widely distributed at lower altitudes.

Conclusion

Conservation

Of the local endemic Oxytropis in the Balkans, we have no information on the status of the subspe-cies of O. dinarica, O. halleri subsp. korabensis or O. prenja, although a website classifies the latter spe-cies as Vulnerable. Oxytropis urumovii is localised to the northern marble part of the Bulgarian Pirin Mts, above 2000 m. Although it is not locally frequent, the population occupies a large territory, part of it with-in a protected region, and in our judgement its sta-tus is not currently under threat. Its IUCN category is Vulnerable according to the Red Data Lists of Vascular Plants and Fungi in Bulgaria (unpubl.).

The only known locality for O. kozhuharovii lies in the northernmost part of the Bulgarian Pirin Mts, known as the Yavorov anticline. Plants were limited to an area of steep rough grassland on marble. A total of about three thousand mature plants were discov-ered after extensive searches but they occupy a ter-ritory of only about 200 m2. Some were of consider-able size and thought to be many decades old. Only a few seedlings or juvenile plants could be found. Although large amounts of seed are apparently set, regeneration appears to be poor. The locality is very remote, and for the moment rarely visited by man, and not heavily grazed by domesticated animals, but might be threatened by land-slips or avalanches. Its


IUCN category is Critically Endangered (unpubl.). The real threat is the current active development of this mountain as a huge ski resort. O. kozhuha-rovii is a distinctive species which must be regard-ed as amongst the most threatened plants in Europe. There is an urgent need to promote its regeneration, perhaps by raising seed in cultivation and planting out seedlings in exclosures to protect them from grazing by wild animals.

Acknowledgements. This work was undertaken under a Royal Society of London Joint Project with Eastern Europe Grant. We are grateful to Dr K. Kostov, Mr G. Matveev and Mr D. Angelov for the help in the field study. Our thanks also go to Mrs V. Tzvetanova for the technical help.

References

Andreev, N. 1981. Reports. – In: Löve, Á. (ed.), Chromosome number reports LXX. – Taxon, 30(1): 74-75.

Andreev, N. 1989. Flora structure of the biosphere reserve “Bayovi doupki – Dzhindzhiritsa” – quantitative parameters, ecobiologi-cal characteristics, genesis status and conservation. Habilitation labor, Inst. Bot., Bulg. Acad. Sci., Sofia, p. 253-255. (in Bulgarian, unpubl.).

Black, W.C. 1995. FORTRAN Programs for the Analysis of RAPD-PCR Markers in Populations. Colorado State Univ., Ft. Collins, CO 80523.

Bottema, S. 1974. Late Quaternary Vegetation History of Northwestern Greece. PhD Thesis. Rijksuniversiteit Groningen, Groningen.

Bottema, S. 1979. Pollen analytical investigations in Thessaly (Greece). – Palaeohistoria, 21: 19-40.

Bozilova, E., Tonkov, S, & Panovska, H. 1989. Palynological investigation in the Kupena National Reserve, West Rhodopes. – Geographica Rhodopica 1. Pp. 186-190. Sofia Univ. St. Kliment Ohridski Press, Sofia.

Bozilova, E. & Tonkov, S. 2000. Pollen from lake Sedmo Rilsko reveals southeast European postglacial vegetation in the highest mountain area of the Balkans. – New Phytol., 148: 315-325.

Bozilova, E. 1996. Bulgaria. Type region BG-e, Vitosha, Rila, Pirin and Rhodope Mountains. – In: Berglund, B. & al. (eds), Palaeoecological events during the last 15 000 years. Pp. 719-721. John Wiley & Sons, Chichester.

Chrtek, J. & Chrtkova, A. 1983. Bemerkungen zu einigen balkanischen Oxytropis-arten. – Folia Geobot. Phytotax., 8: 309-320.

Chung, M., Gelembiuk, G. & Givnish, T.J. 2004. Population genet-ics and phylogeography of endangered Oxytropis campestris var. chartacea and relatives: arctic-alpine disjuncts in eastern North America. – Molec. Ecol., 13(12): 3657-3673.

Diklich, N. 1972. Oxytropis DC. – In: Josifović, M. (ed.), Flore de la Republique Socialiste de Serbie. Vol. 4, pp. 301-307. Acad. Serbe Sci. & Arts, Belgrade (in Serbian).

Felsenstein, J. 1993. PHYLIP (Phylogeny Inference Package) Version 3.5c. Distributed by the author. Dept. of Genetics, Univ. of Washington, Seattle.

Filatov, D.A. 2002. ProSeq: A software for preparation and evolu-tionary analysis of DNA sequence data sets. – Molec. Ecol. Notes, 2: 621-624.

Florschütz, F., Menéndez Amor, J. & Wijmstra, T.A. 1971. Palynology of a thick Quaternary successions in southern Spain. – Palaeogeogr. Palaeoclimatol. Palaeoecol., 10: 233-264.

Greuter, W., Burdet, H.M. & Long, G. (eds.). 1989. Med-Checklist. A Critical Inventory of Vascular Plants of the Circum-mediterranean Countries. Vol. 4. Dicotyledones (Lauraceae-Rhamnaceae). Conservatoire et Jardin Botanique, Med-Checklist trust

Hayek, A. 1927. Prodromus Florae Peninsulae Balcanucae. Band 1. – Repert. Spec. Nov. Regni Veg. Beih., 30(1): 961-1193.

Heywood, V.H. 1967. Plant Taxonomy. Edward Arnold LTD.

Kožuharov, S. 1976. Oxytropis DC. – In: Jordanov, D. (ed.), Fl. Reipubl. Popularis Bulgaricae. Vol. 6, pp. 177-181. In Aedibus Acad. Sci. Bulgaricae, Serdicae (in Bulgarian).

Kruscheva, R. 1986. Reports. – In: Löve, A. (ed.), IOPB Chromosome number reports XCII. – Taxon, 35(3): 613.

Küpfer, P. 1974. Recherches sur les liens de parenté entre la flore orophile des Alpes et celle des Pyrénées. – Boissiera, 23: 12-322.

Lang, G. 1994. Quartare Vegetationgeschichte Europas. Gustav Fisher Verlag, Jena.

Leins, P. & Merxmüller, H. 1968. Oxytropis DC. – In: Tutin, T.G. & al. (eds), Flora Europaea. Vol. 2, pp.124-126. Cambridge Univ. Press, Cambridge.

Page, R.D.M. 1996. TREEVIEW: An application to display phylo-genetic trees on personal computers. – Computer Applic. Biosci., 12: 357-358.

Pavlova, D. 1996. Reports (762–766). – In: Kamari, G., Felber, F. & Garbari, F. (eds), Mediterranean chromosome number re-ports – 6. – Fl. Medit., 6: 323-326.

Pavlova, D., Dimitrov, D. & Nikolova, M. 1999. Oxytropis kozhu-harovii (Fabaceae), a new species from Bulgaria. – Willdenowia, 29: 69-74.

Sanderson, M.J. & Wojciechowski, M.F. 1996. Diversification rates in a temperate legume clade: are there ‘so many’ species of Astragalus (Fabaceae)? – Amer. J. Bot., 83: 1488-1502.

Schönswetter, P., Tribsch, A. & Niklfeld, H. 2004. Amplified frag-ment length polymorphism (AFLP) reveals no genetic divergence of the Eastern Alpine endemic Oxytropis campestris subsp. tiro-liensis (Fabaceae) from widespread subsp. campestris. – Plant Syst. Evol., 244, 245-255.

Strid, A. (ed.). 1986. Mountain Flora of Greece.Vol. 1. Cambridge Univ. Press, Cambridge.

Taberlet, P., Gielly, L., Pautou, G., & Bouvet, J. 1991. Universal primers for amplification of three non-coding regions of chlo-roplast DNA. – Plant Molec. Biol., 17: 1105-1109.


Taschetto, O.M. & Pagliarini, M.S. 2004. Meiotic behaviour in tetraploid populations of Pfaffia tuberosa (Amaranthaceae). – Acta. Bot. Croat., 63(1): 17-24.

Tonkov, S., Panovska, H., Possnert, G. & Bozilova, E. 2002. The Holocene vegetation history of Northern Pirin Mountain, SW Bulgaria: pollen analysis and radiocarbon dating of a core from lake Ribno Banderishko. –Holocene, 12: 201-210.

Tonkov, S., Possnert, G. & Bozilova, E. 2006. The Late Glacial vegetation and radiocarbon dating of lake Trilistnika, Rila Mts (Bulgaria. – Veget. Hist. & Archaeobot., 16(1): 15-22.

Tzedakis, P.C. 1994. Vegetation change through glacial-interglacial cycles: a long pollen sequence perspective. – Philos. Trans. Roy. Soc. London, B345: 403-432.

Tzedakis, P.C. 1999. The last climatic cycle at Kopais, central Greece. – J. Geol. Soc. London, 155: 425-434.

Velchev, A. & Kenderova, R. 1994. Some opinions about the Pleistocene and Holocene development of the Mozgovitza River valley. – God. Sofiisk. Univ. “St. Kliment Ohridski”, Biol. Fak., 2. Bot., 85: 29-42 (in Bulgarian).

Velčev, V., Kožurahov, S. & Ančev, M. (eds). 1992. Atlas of the Endemic Plants in Bulgaria. Publishing Hhouse Bulg. Acad. Sci., Sofia (in Bulgarian).

Weising, K., Nybom, H., Wolff, K., & Meyer, W. 1995. DNA Fingerprinting in Plants and Fungi. CRC Press, Boca Raton.

Wettstein, R. 1892. Beitrag zur Flora Albaniens. – Biblioth. Bot., 26: 1-103.

Wijmstra, T.A. 1969. Palynology of the first 30 metres of a 120 m deep section in northern Greece. – Acta Bot. Neerl., 18: 511-527.

White, T.J., Bruns, T., Lee, S. & Taylor, J.W. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for

phylogenetics. – In: Innis, M.A. & al. (eds), PCR Protocols: A Guide to Methods and Applications. Pp. 315-322. Acad. Press, Inc., New York.

Wojciechowski, M.F., Sanderson, M.J., Baldwin, B.G. & Donoghue, M.J. 1993. Monophyly of aneuploid Astragalus (Fabaceae): evidence from nuclear ribosomal DNA internal transcribed spacer sequences. – Amer. J. Bot., 80: 711-722.

Yurtsev, B.A. 1994. Floristic division of the Arctic. – J. Veg. Sci., 5(6): 765-776.

Yurtsev, B.A. (ed.). 1986. Arctic Flora of the USSR. Vol. 9, part 2. Nauka, Leningrad (in Russian).

Yurtsev, B.A. 1988. A new species of the genus Oxytropis from the section Baicalia (Fabaceae) and the problem of continental relicts in the flora of Wrangel Island. – Bot. Zhurn. (Moscow & Leningrad), 73(12): 1649-1661 (in Russian).

Yurtsev, B.A. 1997. Analysis of evolutionary differentiation in key arctic-alpine taxa: Dryas, Oxytropis sect. Arctobia and Taraxacum sect. Arctica. – Opera Bot., 132: 27-37.

Yurtsev, B. A. 1999. Survey of arctic legumes with emphasis on the species concept in Oxytropis. – Norske Vidensk. – Akad.. Mat.-Naturvidensk. Kl., Avh., 38: 295-318.

Yurtsev, B.A., Katenin, A.E. & Rezvanova, G.S. 1994. Three local floras of carbonate landscapes in the northeastern Chukchee Peninsula. – Bot. Zhurn. (Moscow & Leningrad), 79(1): 34-46 (in Russian).

Yurtsev, B.A., Tolmachev, A.I. & Rebristaya, O.V. 1978. Floristic delimitation and division of the Arctic. – In: Yurtsev, B.A. (ed.), Arctic Floristic Region. Pp. 9-104. Nauka, Leningrad (in Russian).

Zagorchev, I. 1995. Pirin. Geological Guidebook. Acad. Publishing House ‘Marin Drinov’, Sofia.

two rare oxytropis species (fabaceae) endemic to the pirin

Documents